Anthropic cut off access to its Mythos 5 and Fable 5 models after the Trump administration ordered it to block foreign access. The Pentagon’s chief information officer backed the move, while AI researchers—including Gary Marcus and Yann LeCun—criticized the de
For Dario Amodei, the warning was supposed to land like a safety briefing. Instead, it collided with government action—fast enough to make the AI industry feel like it was watching a fuse get lit.
In an essay published this month. the Anthropic CEO warned that the power of AI “has become undeniable.” He pointed to Anthropic’s latest model. Mythos. saying it presents “very real risks” to cybersecurity. the financial sector. critical infrastructure. and national security. He argued the situation calls for stronger government intervention.
On Friday, that intervention arrived in a form that immediately changed how Anthropic’s products reach the world. Anthropic cut off access to its Mythos 5 and Fable 5 after it said the Trump administration ordered it to block foreign access to the models. The move drew sharp support from the Pentagon’s chief information officer. who posted on X that. “Some things are simply more important than revenue cycles. clickbait. and pre-IPO valuation.”.
The shutdown sent shockwaves through the tech world. Gary Marcus, an AI researcher and a longtime skeptic, called the Trump administration’s move “wildly overdramatic and also counterproductive.”
That clash wasn’t only about timing—it was also about who to blame. Yann LeCunn, one of AI’s best-known figures, blamed Amodei directly on X, writing: “Dario Amodei’s ridiculous fear mongering about Mythos/Fable (and AI in general) finally pays off,” and adding, “One reaps what one sows.”
Amodei’s critics have heard this argument before. He is known for positioning himself as the AI industry’s “adult in the room.” Before founding Anthropic. he was an integral researcher at OpenAI. where he left over concerns that OpenAI CEO Sam Altman prioritized releasing products over ensuring their safety.
Since splitting off, Amodei has repeatedly argued that the systems his company is building and releasing to the general public—alongside companies such as OpenAI, Google, Meta, and Amazon—could upend life as we know it, “for good as well as bad.”
He also forecast major disruption: Amodei previously warned that AI would eliminate half of entry-level white-collar jobs and cause unemployment to soar to levels not seen since the COVID-19 pandemic. and before that. the Great Recession in 2008. More recently. he softened his tone on jobs as Anthropic prepares to go public. but his company hasn’t backed away from broader safety concerns.
In his June essay, Amodei wrote: “The cyber risks that Mythos-class models present will not be the last that we must face.” He added that he believes “biological risks may soon follow,” and that “serious AI autonomy risks may not be far behind.”
He also claimed lawmakers’ concerns were out of step with AI’s rapid progress. Earlier this month, Anthropic called for a temporary halt to the development of frontier AI models. In the paper. the company cautioned that the latest models are getting closer to improving themselves. saying this “might increase the risks of humans losing control over AI systems.”.
Anthropic wrote that “it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology.” Friday’s move—at least for Anthropic—turned that plea into something immediate and concrete.
By cutting off access to Mythos 5 and Fable 5 after the Trump administration ordered a block on foreign access. Anthropic effectively accepted that at least one kind of pause can be imposed from the outside. The question now for investors. users. and regulators alike is whether the shutdown reflects a narrow cybersecurity and national-security judgment—or a broader shift in how frontier models are allowed to travel beyond U.S. borders.
Anthropic Dario Amodei Mythos 5 Fable 5 Trump administration Pentagon chief information officer cybersecurity critical infrastructure national security AI regulation frontier AI halt
So they shut it down but for “national security” like… ok sure.
I saw the headline and immediately thought it was just politics. Pentagon supporting it sounds like they’re trying to control AI access, not fix cybersecurity. But then again who even knows what Mythos/Fable does lol.
“Foreign access” sounds vague. Like does that mean they blocked China or Russia or just anyone overseas? Also the article says it’s cybersecurity risks in finance and critical infrastructure but that feels like a stretch without examples. Gary Marcus calling it overdramatic is fair though… government always waits until stuff blows up.
This is what happens when some guy (Amodei?) is like the adult in the room and everyone else is like “fear mongering.” I swear every AI story turns into the same thing where someone shuts off models and calls it safety, but then they’re really just flexing power. If the Pentagon says “more important than revenue cycles” then that’s basically admitting it was about money in the first place. Also the part about “watching a fuse get lit” is so dramatic, yet they acted fast which makes it even weirder.